Re: IMA and ETXTBSY

[Mimi Zohar]

hooanon05@xxxxxxxxxxx wrote on 06/18/2009 01:32:58 AM:

> After the failure of open(2) with ETXTBSY, IMA produces wrong messages.
> Here is a simple test script.
> 
> dir=/tmp
> cp /bin/sleep $dir
> ls -l $dir/sleep
> stat -f $dir
> $dir/sleep 3 & sleep 1
> echo | /bin/dd of=$dir/sleep conv=notrunc && false
> wait
> rm $dir/sleep
> 
> 
> The result.
> + cp /bin/sleep /tmp
> + ls -l /tmp/sleep
> -rwxr-xr-x 1 jro jro 23504 Jun 18 14:26 /tmp/sleep
> + stat -f /tmp
>   File: "/tmp"
>     ID: 4e22c34ea8177d80 Namelen: 255     Type: ext2/ext3
> Block size: 1024       Fundamental block size: 1024
> Blocks: Total: 918322     Free: 918281     Available: 869285
> Inodes: Total: 490560     Free: 490545
> + /tmp/sleep 3
> + sleep 1
> + echo
> + /bin/dd of=/tmp/sleep conv=notrunc
> /bin/dd: opening `/tmp/sleep': Text file busy
> + wait
> + rm /tmp/sleep
> $ iint_free: writecount: 1
> iint_free: opencount: 1
> 
> 
> J. R. Okajima

The problem is that IMA measures the file and increments the 
open/read/write counters before the file is actually opened,
but if the open fails it doesn't decrement the counters.  I'll
post a patch. Thanks for bringing this to my attention.

Mimi
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/